Wyszukiwarka
Filtry
wszystkich: 3255
-
Katalog
- Publikacje 2893 wyników po odfiltrowaniu
- Czasopisma 17 wyników po odfiltrowaniu
- Konferencje 14 wyników po odfiltrowaniu
- Osoby 43 wyników po odfiltrowaniu
- Wynalazki 1 wyników po odfiltrowaniu
- Projekty 3 wyników po odfiltrowaniu
- Kursy Online 68 wyników po odfiltrowaniu
- Wydarzenia 4 wyników po odfiltrowaniu
- Dane Badawcze 212 wyników po odfiltrowaniu
wyświetlamy 1000 najlepszych wyników Pomoc
Wyniki wyszukiwania dla: PRE-TRAINED MODELS
-
Comparison of Language Models Trained on Written Texts and Speech Transcripts in the Context of Automatic Speech Recognition
Publikacja -
Training of Deep Learning Models Using Synthetic Datasets
PublikacjaIn order to solve increasingly complex problems, the complexity of Deep Neural Networks also needs to be constantly increased, and therefore training such networks requires more and more data. Unfortunately, obtaining such massive real world training data to optimize neural networks parameters is a challenging and time-consuming task. To solve this problem, we propose an easy-touse and general approach to training deep learning...
-
MagMax: Leveraging Model Merging for Seamless Continual Learning
PublikacjaThis paper introduces a continual learning approach named MagMax, which utilizes model merging to enable large pre-trained models to continuously learn from new data without forgetting previously acquired knowledge. Distinct from traditional continual learning methods that aim to reduce forgetting during task training, MagMax combines sequential fine-tuning with a maximum magnitude weight selection for effective knowledge integration...
-
Assessing the attractiveness of human face based on machine learning
PublikacjaThe attractiveness of the face plays an important role in everyday life, especially in the modern world where social media and the Internet surround us. In this study, an attempt to assess the attractiveness of a face by machine learning is shown. Attractiveness is determined by three deep models whose sum of predictions is the final score. Two annotated datasets available in the literature are employed for training and testing...
-
Melanoma skin cancer detection using mask-RCNN with modified GRU model
PublikacjaIntroduction: Melanoma Skin Cancer (MSC) is a type of cancer in the human body; therefore, early disease diagnosis is essential for reducing the mortality rate. However, dermoscopic image analysis poses challenges due to factors such as color illumination, light reflections, and the varying sizes and shapes of lesions. To overcome these challenges, an automated framework is proposed in this manuscript. Methods: Initially, dermoscopic...
-
Justyna Signerska-Rynkowska dr inż.
OsobySince 2016 assistant professor at Gdańsk University of Technology, Faculty of Applied Physics and Mathematics, Division of Differential Equations and Mathematics Applications 2021-2024 visiting assistant professor in Dioscuri Centre in Topological Data Analysis (Institute of Mathematics of the Polish Academy of Sciences, IMPAN) Since 2020 - Principal Investigator in "SONATA" grant “Challenges of low-dimensional...
-
A Mammography Data Management Application for Federated Learning
PublikacjaThis study aimed to develop and assess an application designed to enhance the management of a local client database consisting of mammographic images with a focus on ensuring that images are suitably and uniformly prepared for federated learning applications. The application supports a comprehensive approach, starting with a versatile image-loading function that supports DICOM files from various medical imaging devices and settings....
-
News that Moves the Market: DSEX-News Dataset for Forecasting DSE Using BERT
PublikacjaStock market is a complex and dynamic industry that has always presented challenges for stakeholders and investors due to its unpredictable nature. This unpredictability motivates the need for more accurate prediction models. Traditional prediction models have limitations in handling the dynamic nature of the stock market. Additionally, previous methods have used less relevant data, leading to suboptimal performance. This study...
-
Selection of an artificial pre-training neural network for the classification of inland vessels based on their images
PublikacjaArtificial neural networks (ANN) are the most commonly used algorithms for image classification problems. An image classifier takes an image or video as input and classifies it into one of the possible categories that it was trained to identify. They are applied in various areas such as security, defense, healthcare, biology, forensics, communication, etc. There is no need to create one’s own ANN because there are several pre-trained...
-
Deep Instance Segmentation of Laboratory Animals in Thermal Images
PublikacjaIn this paper we focus on the role of deep instance segmentation of laboratory rodents in thermal images. Thermal imaging is very suitable to observe the behaviour of laboratory animals, especially in low light conditions. It is an non-intrusive method allowing to monitor the activity of animals and potentially observe some physiological changes expressed in dynamic thermal patterns. The analysis of the recorded sequence of thermal...
-
Klasyfikator SVM w zastosowaniu do synchronizacji sygnału OFDM zniekształconego przez kanał wielodrogowy
PublikacjaW pracy przedstawiono analizę przydatności klasyfikatora SVM bazującego na uczeniu maszynowym do estymacji przesunięcia czasowego odebranego symbolu OFDM. Przedstawione wyniki wykazują, że ten klasyfikator potrafi zapewnić synchronizację dla różnych kanałów wielodrogowych o wysokim poziomie szumu. Eksperymenty przeprowadzone w Matlabie z użyciem modeli modulatora i demodulatora wykazały, że w większości przypadków klasyfikator...
-
Real-Time Skin Quality Assessment System
PublikacjaThis thesis presents a real-time skin assessment system with the main aim of detecting inflammatory acne lesions and tracking skin conditions. A facial acne lesion detection algorithm was developed, using a pre-trained YOLOv8 model for lesion detection and a Mediapipe detector for face detection. The proposed solution aims to create a system used in smart mirrors to help users self-monitor and monitor their skin condition in real...
-
LSA Is not Dead: Improving Results of Domain-Specific Information Retrieval System Using Stack Overflow Questions Tags
PublikacjaThe paper presents the approach to using tags from Stack Overflow questions as a data source in the process of building domain-specific unsupervised term embeddings. Using a huge dataset of Stack Overflow posts, our solution employs the LSA algorithm to learn latent representations of information technology terms. The paper also presents the Teamy.ai system, currently developed by Scalac company, which serves as a platform that...
-
Potential and Use of the Googlenet Ann for the Purposes of Inland Water Ships Classification
PublikacjaThis article presents an analysis of the possibilities of using the pre-degraded GoogLeNet artificial neural network to classify inland vessels. Inland water authorities monitor the intensity of the vessels via CCTV. Such classification seems to be an improvement in their statutory tasks. The automatic classification of the inland vessels from video recording is a one of the main objectives of the Automatic Ship Recognition and...
-
Poprawa jakości klasyfikacji głębokich sieci neuronowych poprzez optymalizację ich struktury i dwuetapowy proces uczenia
PublikacjaW pracy doktorskiej podjęto problem realizacji algorytmów głębokiego uczenia w warunkach deficytu danych uczących. Głównym celem było opracowanie podejścia optymalizującego strukturę sieci neuronowej oraz zastosowanie uczeniu dwuetapowym, w celu uzyskania mniejszych struktur, zachowując przy tym dokładności. Proponowane rozwiązania poddano testom na zadaniu klasyfikacji znamion skórnych na znamiona złośliwe i łagodne. W pierwszym...
-
Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning
PublikacjaExemplar-Free Class Incremental Learning (EFCIL) tackles the problem of training a model on a sequence of tasks without access to past data. Existing state-of-the-art methods represent classes as Gaussian distributions in the feature extractor's latent space, enabling Bayes classification or training the classifier by replaying pseudo features. However, we identify two critical issues that compromise their efficacy when the feature...
-
Wymiary klimatu pracy zespołowej
PublikacjaArtykuł prezentuje koncepcje klimatu pracy zespołowej jako determinanty zespołowej efektywności i innowacyjności. Zaprezentowano sześć modeli klimatu pracy zespołowej i przeanalizowano główne wymiary klimatu, opracowując na tej podstawie nowy, 5-wymiarowy model, który obejmuje: wsparcie oparte na zaufaniu, popieranie innowacji,odpowiedzialność, wizja i motywacja. Pokazano również kilka narzędzi do pomiaru klimatu pracy zespołowej.
-
Adaptacyjny system sterowania ruchem drogowym
PublikacjaAdaptacyjny system sterowania ruchem drogowym to rodzaj systemu sterowania, który dynamicznie, w czasie rzeczywistym, dostosowuje swoje parametry w oparciu o bieżące warunki ruchu drogowego. Celem niniejszej rozprawy jest sprawdzenie wpływu wybranych cech systemu, zbudowanego w oparciu o zaprojektowane i zbudowane z udziałem autora inteligentne znaki drogowe, na wybrane parametry mające wpływ na bezpieczeństwo i płynność ruchu....
-
Machine Learning and Deep Learning Methods for Fast and Accurate Assessment of Transthoracic Echocardiogram Image Quality
PublikacjaHigh-quality echocardiogram images are the cornerstone of accurate and reliable measurements of the heart. Therefore, this study aimed to develop, validate and compare machine learning and deep learning algorithms for accurate and automated assessment of transthoracic echocardiogram image quality. In total, 4090 single-frame two-dimensional transthoracic echocardiogram...
-
OrphaGPT: An Adapted Large Language Model for Orphan Diseases Classification
PublikacjaOrphan diseases (OD) represent a category of rare conditions that affect only a relatively small number of individuals. These conditions are often neglected in research due to the challenges posed by their scarcity, making medical advancements difficult. Then, the ever-evolving medical research and diagnosis landscape calls for more attention and innovative approaches to address the complex challenges of rare diseases and OD. Pre-trained...
-
Attention-Based Deep Learning System for Classification of Breast Lesions—Multimodal, Weakly Supervised Approach
PublikacjaBreast cancer is the most frequent female cancer, with a considerable disease burden and high mortality. Early diagnosis with screening mammography might be facilitated by automated systems supported by deep learning artificial intelligence. We propose a model based on a weakly supervised Clustering-constrained Attention Multiple Instance Learning (CLAM) classifier able to train under data scarcity effectively. We used a private...
-
Divide and not forget: Ensemble of selectively trained experts in Continual Learning
PublikacjaClass-incremental learning is becoming more popular as it helps models widen their applicability while not forgetting what they already know. A trend in this area is to use a mixture-of-expert technique, where different models work together to solve the task. However, the experts are usually trained all at once using whole task data, which makes them all prone to forgetting and increasing computational burden. To address this limitation,...
-
Neural Network Subgraphs Correlation with Trained Model Accuracy
PublikacjaNeural Architecture Search (NAS) is a computationally demanding process of finding optimal neural network architecture for a given task. Conceptually, NAS comprises applying a search strategy on a predefined search space accompanied by a performance evaluation method. The design of search space alone is expected to substantially impact NAS efficiency. We consider neural networks as graphs and find a correlation between the presence...
-
Neural network agents trained by declarative programming tutors
PublikacjaThis paper presents an experimental study on the development of a neural network-based agent, trained using data generated using declarative programming. The focus of the study is the application of various agents to solve the classic logic task – The Wumpus World. The paper evaluates the effectiveness of neural-based agents across different map configurations, offering a comparative analysis to underline the strengths and limitations...
-
Obtaining a Well-Trained Artificial Intelligence Algorithm from Cross-Validation in Endoscopy
PublikacjaThe article shortly discusses endoscopic video analysis problems and artificial intelligence algorithms supporting it. The most common method of efficiency testing of these algorithms is to perform intensive cross-validation. This allows for accurately evaluate their performance of generalization. One of the main problems of this procedure is that there is no simple and universal way of obtaining a specific instance of a well-trained...
-
TNFA expression level changes observed in response to the Wingate Anaerobic Test in non-trained and trained individuals
Publikacja -
Models in Spatial Development
Kursy Online -
Preeclampsia Risk Prediction Using Machine Learning Methods Trained on Synthetic Data
PublikacjaThis paper describes a research study that investigates the use of machine learning algorithms on synthetic data to classify the risk of developing preeclampsia by pregnant women. Synthetic datasets were generated based on parameter distributions from three real patient studies. Four models were compared: XGBoost, Support Vector Machine (SVM), Random Forest, and Explainable Boosting Machines (EBM). The study found that the XGBoost...
-
STOCHASTIC MODELS
Czasopisma -
GRAPHICAL MODELS
Czasopisma -
Pre-oxidation of porous ferritic Fe22Cr alloys for lifespan extension at high-temperature
PublikacjaPre-oxidation of porous ferritic Fe22Cr alloys was extensively studied in this paper. Weight gain measurements and SEM analysis revealed that pre-oxidation performed at 900◦C for 40 min increased the lifespan of the alloy. A Cr evaporation study did not disclose any significant influence of the pre-oxidation process on the Cr content in the alloy. For a more detailed assessment, TEM imaging and X-ray tomography measurements of...
-
Pre-swirl energy saving device in marine application
PublikacjaThis paper covers topics of energy saving device (ESD) with application to marine propulsors. The form of ESD, considered in this paper, consists of fixed lifting foils mounted in front of the screw propeller (the pre-swirl stator/guide vanes). An algorithm for designing propulsion systems, consisting of guide vanes and screw propeller, is presented. The proposed method relies on hybrid lifting line (guide vanes)-lifting surface...
-
Accelerometer signal pre-processing influence on human activity recognition
PublikacjaA study of data pre-processing influence on accelerometer-based human activity recognition algorithms is presented. The frequency band used to filter-out the accelerometer signals and the number of accelerometers involved were considered in terms of their influence on the recognition accuracy.
-
Hydrogen degradation of pre-oxidized zirconium alloys
PublikacjaThe presence of the oxide layers on Zr alloys may retard or enhance the hydrogen entry and material degradation, depending on the layer features. This research has been aimed to determine the effects of pre-oxidation of the Zircaloy-2 alloy at a different temperature on hydrogen degradation. The specimens were oxidised in laboratory air at 350°C, 700°C, and 900°C. After, some samples were tensed at 10-5 strain rate and simultaneously...
-
Models in spatial development 2024/25
Kursy Online -
Malignant neoplasm arising from pre-existing spiradenoma - Male, 64 - Tissue image [912072957516291]
Dane BadawczeThis is the histopathological image of SKIN tissue sample obtained in Medical University Gdańsk and deposited in ZMDL-GUMED. The sample image was taken using: Pannoramic 250 3DHistech slide scanner (20x magnification) and saved to DICOM format.
-
Optimization of Saccharification Conditions of Lignocellulosic Biomass under Alkaline Pre-Treatment and Enzymatic Hydrolysis
PublikacjaPre-treatment is a significant step in the production of second-generation biofuels from waste lignocellulosic materials. Obtaining biofuels as a result of fermentation processes requires appropriate pre-treatment conditions ensuring the highest possible degree of saccharification of the feed material. An influence of the following process parameters were investigated for alkaline pre-treatment of Salix viminalis L.: catalyst concentration...
-
Journal of Pre-Clinical and Clinical Research
Czasopisma -
Comparison of image pre-processing methods in liver segmentation task
PublikacjaAutomatic liver segmentation of Computed Tomography (CT) images is becoming increasingly important. Although there are many publications in this field there is little explanation why certain pre-processing methods were utilised. This paper presents a comparison of the commonly used approach of Hounsfield Units (HU) windowing, histogram equalisation, and a combination of these methods to try to ascertain what are the differences...
-
Hydrogen Production from Energy Poplar Preceded by MEA Pre-Treatment and Enzymatic Hydrolysis
PublikacjaThe need to pre-treat lignocellulosic biomass prior to dark fermentation results primarily from the composition of lignocellulose because lignin hinders the processing of hard wood towards useful products. Hence, in this work a two-step approach for the pre-treatment of energy poplar, including alkaline pre-treatment and enzymatic saccharification followed by fermentation has been studied. Monoethanolamine (MEA) was used as the...
-
Malignant neoplasm arising from pre-existing spiradenoma - Male, 63 - Tissue image [9220729551204161]
Dane BadawczeThis is the histopathological image of BRONCHUS AND LUNG tissue sample obtained in Medical University Gdańsk and deposited in ZMDL-GUMED. The sample image was taken using: Pannoramic 250 3DHistech slide scanner (20x magnification) and saved to DICOM format.
-
Effect of modified soybeen oil amount on rheological characteriza-tion of polyurethane pre-polymers
PublikacjaPolyurethanes (Pu’s) are the polymeric materials which have got urethane groups in the structure. The properties of Pu’s depend both on the method of preparation and monomers used. Polyurethanes are produced by two methods known as one step or two step method called as “pre-polymers method”, especially for the case of segmented polyurethanes (SPU’s). These materials are thermoplastic block copolymers of the (AB)n type consisting...
-
Pre-analytical aspects in metabolomics of human biofluids – sample collection, handling, transport, and storage
PublikacjaMetabolomics is the field of omics research that offers valuable insights into the complex composition of biological samples. It has found wide application in clinical diagnostics, disease investigation, therapy prediction, monitoring of treatment efficiency, drug discovery, or in-depth analysis of sample composition. A suitable study design constitutes the fundamental requirements to ensure robust and reliable results from the...
-
Serendipitous Recommendations Through Ontology-Based Contextual Pre-filtering
PublikacjaContext-aware Recommender Systems aim to provide users with better recommendations for their current situation. Although evaluations of recommender systems often focus on accuracy, it is not the only important aspect. Often recommendations are overspecialized, i.e. all of the same kind. To deal with this problem, other properties can be considered, such as serendipity. In this paper, we study how an ontology-based and context-aware...
-
Fermentative Conversion of Two-Step Pre-Treated Lignocellulosic Biomass to Hydrogen
PublikacjaFermentative hydrogen production via dark fermentation with the application of lignocellulosic biomass requires a multistep pre-treatment procedure, due to the complexed structure of the raw material. Hence, the comparison of the hydrogen productivity potential of different lignocellulosic materials (LCMs) in relation to the lignocellulosic biomass composition is often considered as an interesting field of research. In this study,...
-
Journal of Pre-Raphaelite Studies
Czasopisma -
An Ontology-based Contextual Pre-filtering Technique for Recommender Systems
PublikacjaContext-aware Recommender Systems aim to provide users with the most adequate recommendations for their current situation. However, an exact context obtained from a user could be too specific and may not have enough data for accurate rating prediction. This is known as the data sparsity problem. Moreover, often user preference representation depends on the domain or the specific recommendation approach used. Therefore, a big effort...
-
Influence of accelerometer signal pre-processing and classification method on human activity recognition
PublikacjaA study of data pre-processing influence on accelerometer-based human activity recognition algorithms is presented. The frequency band used to filter-out the accelerometer signals and the number of accelerometers involved were considered in terms of their influence on the recognition accuracy. In the test four methods of classification were used: support vector machine, decision trees, neural network, k-nearest neighbor.
-
Pre-treatment of bio fraction waste prior to fermentation processes
PublikacjaCurrent efforts are taken to increase resource efficiency, close material loops, and improve sustainable waste and by-products management. Thus, networking agro-food by-products and converting them into valuable products completely exhausting the potential of the raw material becomes significant. Model lignocellulosic and starch based biomass were subjected to pre-treatment with the application of acidic compounds, i.e. sulphuric...
-
LCF behavior of 2024AA under uni- and biaxial loading taking into account creep pre-deformation
PublikacjaThis study presents the results of experimental low-cycle fatigue (LCF) tests of aluminum 2024 alloy T3511 temper in uni- and biaxial loading states. Tests were carried out on both the as-received material (hardened extruded rods) and material with different pre-deformation histories. These deformations were carried out in the creep process at 200 °C and 300 °C for two different levels of at each temperature. The pre-deformed material’s...